e2e: add test for cpnp replication

[GiedriusS]

Add e2e test that shows that cpnp replication works. Remove some unused code.

[GiedriusS]

Very weird - writes succeed and I can see the timestamps/values coming in as expected, and the series is in the WAL but I can't query them :/ something is wrong with the writer but I cannot spot anything wrong so far.

[fpetkovski]

I added some logging to the writer and I can also see

Appending sample {__name__="myself"} 1729090293939 1
err: <nil>

Maybe something with the timestamp format?

[fpetkovski]

Also seeing metrics in the ingestor

# HELP prometheus_tsdb_head_samples_appended_total Total number of appended samples.
# TYPE prometheus_tsdb_head_samples_appended_total counter
prometheus_tsdb_head_samples_appended_total{tenant="default-tenant",type="float"} 1
prometheus_tsdb_head_samples_appended_total{tenant="default-tenant",type="histogram"} 0

[fpetkovski]

@GiedriusS changing this line from unsafe.String(&b[0], len(b)) to string(b) makes the test pass. I assume that labels.Copy here only does a shallow copy of labels and actual name and value bytes get lost at some point.

For now we can either remove the unsafe conversion, or we can have a deep copy of labels which re-allocates the name and value strings.

[GiedriusS]

Added a strings.Clone() where it is really necessary. I think it is preferable to always making a copy because creating a new series is a much rarer occurrence than adding new samples to an existing series.

[fpetkovski]

We need to use lset.Range

[fpetkovski]

I'll fix this on my branch.